Detecting the end of the self-write cycle of the AT24CXX

The self-write cycle of AT24CXX is "less than 5ms", so it can be solved by using a delay function to delay 5ms.

Pain point: When writing a delay function, you can use software delay and timer delay. For software delay, you need an oscilloscope to see the exact time. It is very troublesome to use an oscilloscope or not. Moreover, if the main frequency of the microcontroller changes, the value of the software delay needs to be changed. The timer delay takes up a timer and wastes microcontroller resources. In addition, if the AT24CXX suddenly extends the writing time due to its own problems, and this is not taken into account when writing the program, the whole system will be ruined.

Solution: Look at the write timing. After the microcontroller sends a start signal to the AT24CXX and then sends an address data, the AT24CXX will return an ACK signal. If the self-write cycle is not over, it will not return an ACK signal. Therefore, we can observe whether there is this ACK to determine whether the self-write cycle is over. In this way, no matter how long the self-write cycle of the 24CXX is, you can immediately write another data or perform a read operation to improve the efficiency of the program.

The following is a program written using the STM32 library function. The implementation steps are the same as those described in the above solution.

void WaitWriteCycleOver(void)
{
 do
 {
 /* Send start signal */
 I2C_GenerateSTART(I2C1, ENABLE);
 /* Read SR1 register to clear the start flag */
 I2C_ReadRegister(I2C1, I2C_Register_SR1);
 /* Send device address */
 I2C_Send7bitAddress(I2C1, 0xA0, I2C_Direction_Transmitter);
 /* Check whether the device address is sent successfully */
 }while(!(I2C_ReadRegister(I2C1, I2C_Register_SR1) & I2C_FLAG_ADDR));
 
 /* Clear ACK flag */
 I2C_ClearFlag(I2C1, I2C_FLAG_AF);
 /* Send a stop signal */ 
 I2C_GenerateSTOP(I2C1, ENABLE); 
}

After reading this, you may have some questions about the program:

Question 1: Why detect ADDR instead of ACK?

Because we are using hardware I2C, the detection of ACK is done by hardware. If ADDR in SR1 is set, it means that the hardware has detected ACK, otherwise, ACK is not detected.

Question 2: Why do we need to clear ACK alone and not other flags?

Because the ACK bit needs to be cleared by software, the other bits are automatically cleared after reading.